Recovery of micro gas sensor array response from drift effect by means of DWT and feature extraction for discriminating gas mixtures

In the case of semiconductor-type gas sensors, since the reproductivity and selectivity are often declined because of a drift effect, it is hard to extract important features from the micro gas sensor array response to discriminate gas mixtures. In this paper, by using the DWT(Discrete Wavelet Transform) we eliminate the factors which cause the drift and extract the features from the transformed patterns of thermally modulated micro gas sensor array responses. This method makes it possible to minimize the drift effect and to extract important features easily and effectively.